9603839 Onuchic The funnel landscape theory of protein folding will be fully developed. The main idea that emerges from the statistical energy landscape theory is that globally the folding landscape resembles a funnel that is to some extent rugged, i.e. riddled with traps, in which the protein can transiently reside. Detailed simulations of lattice and off-lattice minimalist models will be performed to investigate the details of the folding energy landscape. Using a law of corresponding states, the connection between the landscapes of simple minimalist models and the ones of small fast folding proteins becomes possible. An important feature of this funnel picture is that it enables the determination of the ensemble of configurations that participate in the folding event simply by thermodynamically sampling the configurational space. No kinetic information is necessary. This is essential since full kinetic simulations are impossible for a protein with full atomic details. With the aid of the law of corresponding states and the sampling scheme, simulation of proteins at full atomic details will become possible and one will be able to predict quantitatively the importance of different amino-acids in the folding kinetics of different proteins. The initial protein that will be studied is a three-helix bundle protein (fragment B of Staphylococcal protein A) because of its simplicity (fully alpha-helical and preliminary simulations already exist) but it will be followed by investigations of other fast folding proteins, such as the monomeric lambda-repressor and chymotrypsin inhibitor 2, for which many more site-directed mutagenesis experiments exploring folding stability and folding kinetics are available. To understand how proteins fold is of great importance in biology. To perform its biological function, a protein can not be thought of simply as a linear string of amino acids; its activity depends on its three dimensional structure and dynamics. How can one predict the protein structu re by simply knowing its sequence? A robust productive method would aid the understanding of biological processes at the molecular level by providing deeper insights into the structure/function relationship in biomolecular activity. Although this question has haunted scientists for several decades, very little progress has been made until recent years. Where knowledge is being gained, it arises from the integration of information from physics, chemistry, molecular biology and computer science. Achieving this integration towards the understanding of the folding mechanism is the goal of this work.
